MODELING AND FORECASTING EMERGENCY DEPARTMENTS’ PATIENTS FLOW, IN GRAND-EST REGION

Overview & Synopsis

Hospital tensions in the French emergency departments (EDs) have increased last years and are widely publicized. Organizational, structural or cyclical factors can explain this situation and encourage an understanding and anticipation of the flows of patients in the EDs. The modeling of the flow of passage in hospital departments, including EDs, has been the subject of numerous publications in recent years, in France and internationally. In this paper, we propose a generalized additive model of daily flows, incorporating exogenous information (such as school holiday periods), in which trend is a linear combination of orthogonal polynomials function and seasonality is a combination of harmonics learned from the data. This model is adjusted individually for each ED of the French Grand Est area from the series of passages provided by the association Est-RESCUE. The flow forecasting performances of the proposed model are evaluated and compared to those of the classical SNAIVE, SARIMA and PROPHET.